Role of bacterial efflux pumps in antibiotic resistance, virulence, and strategies to discover novel efflux pump inhibitors

Abstract

The problem of antibiotic resistance among pathogenic bacteria has reached a crisis level. The treatment options against infections caused by multiple drug-resistant bacteria are shrinking gradually. The current pace of the discovery of new antibacterial entities is lagging behind the rate of development of new resistance. Efflux pumps play a central role in making a bacterium resistant to multiple antibiotics due to their ability to expel a wide range of structurally diverse compounds. Besides providing an escape from antibacterial compounds, efflux pumps are also involved in bacterial stress response, virulence, biofilm formation, and altering host physiology. Efflux pumps are unique yet challenging targets for the discovery of novel efflux pump inhibitors (EPIs). EPIs could help rejuvenate our currently dried pipeline of antibacterial drug discovery. The current article highlights the recent developments in the field of efflux pumps, challenges faced during the development of EPIs and potential approaches for their development. Additionally, this review highlights the utility of resources such as natural products and machine learning to expand our EPIs arsenal using these latest technologies.

Full text

1
Role of bacterial eux pumps in antibiotic resistance, virulence,
and strategies to discover novel eux pumpinhibitors
AmitGaurav, PerwezBakht, MahakSaini, ShivamPandey and RanjanaPathania*
REVIEW
Gaurav etal., Microbiology 2023;169:001333
DOI 10.1099/mic.0.001333
Received 02 February 2023; Accepted 24 April 2023; Published 24 May 2023
Author aliations: 1Department of Biosciences and Bioengineering, Indian Institute of Technology Roorkee, Roorkee, Uttarakhand, India.
*Correspondence: Ranjana Pathania, ranjana. pathania@ bt. iitr. ac. in
Keywords: eux pump; EPI; AcrAB- TolC; antibacterial adjuvants; Antibiotic resistance; machine learning.
Abbreviations: ABC, ATP- binding cassette; ADP, adenosine diphosphate; AMR, antimicrobial resistance; ATP, adenosine triphosphate; EPI, eux pump
inhibitor; MATE, multidrug and toxic compound extrusion; MFS, major facilitator superfamily; NBD, nucleotide binding domain; PACE, proteobacterial
antimicrobial compound eux; RND, resistance nodulation cell division; SMR, small multidrug resistance.
001333 © 2023 The Authors
This is an open- access article distributed under the terms of the Creative Commons Attribution License. The Microbiology Society waived the open access fees for this article.
Abstract
The problem of antibiotic resistance among pathogenic bacteria has reached a crisis level. The treatment options against infec-
tions caused by multiple drug- resistant bacteria are shrinking gradually. The current pace of the discovery of new antibacterial
entities is lagging behind the rate of development of new resistance. Eux pumps play a central role in making a bacterium
resistant to multiple antibiotics due to their ability to expel a wide range of structurally diverse compounds. Besides provid-
ing an escape from antibacterial compounds, eux pumps are also involved in bacterial stress response, virulence, biofilm
formation, and altering host physiology. Eux pumps are unique yet challenging targets for the discovery of novel eux pump
inhibitors (EPIs). EPIs could help rejuvenate our currently dried pipeline of antibacterial drug discovery. The current article
highlights the recent developments in the field of eux pumps, challenges faced during the development of EPIs and potential
approaches for their development. Additionally, this review highlights the utility of resources such as natural products and
machine learning to expand our EPIs arsenal using these latest technologies.
INTRODUCTION
e discovery of antibiotics in the mid- twentieth century transformed medical sciences and consequently enhanced life expec-
tancy across the globe [1]. Antibiotics provide eective infection management, which enables us to deal with complex medical
procedures such as surgeries, organ transplants, cancer therapy, and many more [2]. e joy of this antibiotic era, however, did not
last long due to the simultaneous evolution of antimicrobial resistance (AMR) [3]. In the current era, antibiotic- resistant bacteria
are a global threat to health and livelihoods. ough antibiotic resistance predates the use of antibiotics clinically, inappropriate
and misuse of antibiotics has led to the evolution of new AMR mechanisms in pathogenic bacteria [4–7]. e emergence and
global spread of AMR jeopardize the ability to treat infectious diseases and raises healthcare costs. e 2017 report of the World
Health Organization on antibiotic resistance priority pathogens also highlights the importance of focusing research on these
deadly microbes [8]. Antibiotic resistance in bacteria can be developed through six major mechanisms; (a) limiting the uptake of
the antibiotic by altering the cellular permeability, (b) modifying the antibiotic target site, (c) target site protection, (d) enzymatic
inactivating of the antibiotic, (e) active antibiotic eux pumps, and (f) target bypass (Fig.1) [9].
Antibiotic eux is one of the most common mechanisms of resistance among a wide range of pathogenic bacteria [10, 11].
Eux pumps are transport proteins localized in the cytoplasmic membrane of bacteria that actively translocate the chemical
across the membrane. Eux pumps are involved in the regulation of the internal environment by extruding out the toxic
substances, quorum sensing molecules (autoinducers), biolm formation molecules, and virulence factors of the bacteria
(Fig.2) [12]. Eux pumps can confer heavy metal resistance by exporting metal ions such as Ag2+, Cu2+, Co2+, Zn2+, Cd2+
and Ni
2+
. In Gram- negative bacteria, eux pumps help to reduce not only the cytoplasmic concentration of heavy metal ions
but also the periplasmic concentration; since periplasmic metal ions can re- enter the cytoplasm, multiple eux pumps work
together to evade heavy metal toxicity [13–15]. Eux pumps are categorized as primary and secondary on the basis of the
source of energy they utilize to pump out the substrates. e eux pumps that drive energy from the hydrolysis of ATP to
translocate the substrates across the membrane are dened as the primary eux pumps, whereas those which draw energy
OPEN
ACCESS

2
Gaurav etal., Microbiology 2023;169:001333
from electrochemical gradients formed by protons or ions (proton motive force) are dened as the secondary eux pumps.
Till now, six major families of eux pumps have been found in bacteria, namely, ATP- binding cassette (ABC) superfamily,
major facilitator superfamily (MFS), multidrug and toxic compound extrusion (MATE), resistance nodulation cell division
(RND) family, small multidrug resistance (SMR) family and proteobacterial antimicrobial compound eux (PACE) family
(Fig.3) [16]. Even though there are various types of eux pumps, substrate redundancy exists among all classes of eux
pumps, such that one antibiotic can be exported by several dierent eux pumps and a single eux pump can export
structurally and chemically diverse substrates [17]. Eux pumps play an essential role in dierent stress environments
for bacteria; thus, they can be a promising target for developing new inhibitors to rejuvenate obsolete existing antibiotics.
However, due to the substrate redundancy of eux pumps, the clinical success of eux pump inhibitors is uncertain. is
challenge can be addressed by discovering novel broad- spectrum eux pump inhibitors.
e current review summarizes recent advancements in our understanding of the role of eux pumps in bacterial physiology.
Furthermore, this review will focus on dierent strategies for the development of eux pump inhibitors, such as chemoinformatics
and machine learning.
TYPES OF BACTERIAL EFFLUX SYSTEMS
ATP-binding cassette (ABC) superfamily
e ABC superfamily is a primary eux pump family and it draws energy from the active hydrolysis of ATP to translocate
a wide range of solutes, including drugs, lipids, and sterols across the membrane [18, 19]. All ABC families share a basic
common architecture of two membrane- integral part domains that transverse the membrane six times each (12 transmem-
brane domains [TMDs]) and two nucleotide binding domains (NBDs) where the TMDs are involved in substrate- binding
Fig. 1. Schematic representation of antibiotic resistance mechanisms. Six dierent resistance mechanisms are found in bacteria. (a)Increased eux
by eux systems; (b)reduced uptake due to change in membrane permeability; (c)enzymatic degradation; (d)target site protection; (e)target site
modification and (f)expression of alternative enzymes or o- target sites.

3
Gaurav etal., Microbiology 2023;169:001333
and NBD bind and hydrolyse ATP to make the transport cycle work [16, 20, 21]. ABC extrusion systems are divided into
‘full’ or ‘half’ transporters. In full transporters, a single polypeptide encodes two NBDs and two TMDs. In half transporters,
a single polypeptide encodes for one NBD and one TMD [18, 22]. e ABC family functions according to an 'alternate
access' mechanism between the inward- facing (IF) and the outward- facing (OF) to extrude substrates across the membrane,
but the extent of physical separation between the two NBDs in the inward- facing (IF) and the outward- facing (OF) states
is still unsettled [16]. In addition, a recent study reported on the physical segregation between two NBDs and found that
NBD segregation exclusively decreases from inward to outward facing state, whereas NBD segregation seems larger when
ADP is bound to it [23]. ABC transporters found in human pathogenic bacteria contribute to virulence, pathogenesis, and
multidrug resistance via various mechanisms. ABC transporters have dual functionality, i.e. they can act as an importer or
an exporter. ABC importers aid virulence by acquiring essential nutrients such as peptides, vitamins, amino acids, transi-
tion metals, and osmoprotectants. ABC transporters contribute to virulence by exporting essential molecules involved in
glycoconjugate biosynthesis such as lipopolysaccharides and capsular polysaccharides, as well as by exporting xenobiotics
[24]. In Salmonella enterica serovar Typhimurium, MacAB an eux pump belonging to the ABC family is shown to aid its
survival when exposed to oxidative stress inside the macrophages [25]. Transition metals play critical role in pathogenesis of
several human pathogens [26]. ABC family eux pumps are reported to be involved in export of transition metals in major
human pathogens like Pseudomonas aeruginosa, Listeria monocytogenes, and Mycobacterium tuberculosis [27–29]. e diverse
functionality of ABC transporters makes them an attractive target for antibacterial drug development.
Resistance nodulation cell division (RND) superfamily
e RND superfamily mainly consists of twelve transmembrane helices that are separated by two large loops to form asym-
metric trimers and the outer loops contain binding sites for exported ligands, while the transmembrane domain mainly
functions as a channel for protons to utilise energy for substrates translocation [30]. ough 12 helices are signature of RND
eux pumps, sometimes more than 12 helices have been reported, for example, SecYEG of Escherichia coli has 15 trans-
membrane helices [31]. RND eux pumps are known to work as a trimer (for example HAE [hydrophobic and amphiphilic
eux]-subfamily eux pumps, including AcrB of E. coli, MexB of P. aeruginosa, MtrD of Neisseria gonorrhoeae, CmeB of
Campylobacter jejuni, and HME [heavy metal eux]-subfamily proteins, such as CusA and ZneA), although recent studies
suggest that these transporters may be either dimers (HpnN membrane protein of Burkholderia multivorans) or monomers
(MmpL3 transporter of Mycobacterium smegmatis) [16, 32]. e RND superfamily is the most potent eux pump family
Fig. 2. Schematic representation of biological functions of bacterial eux pumps. Eux pumps play an important role in (a)eux of antibiotics,
(b)biofilm formation, (c)regulation of host physiology, (d)metal resistance, and (e)virulence.

4
Gaurav etal., Microbiology 2023;169:001333
mediating antibiotic resistance in Gram- negative bacteria. e pump is composed of an outer membrane factor (OMF),
a periplasmic adapter protein (PAP) and an inner membrane RND- transporter, and thus also called tripartite resistance-
nodulation- division (RND) eux pumps [33–35]. Furthermore, the PAP is a complex protein comprised of four domains:
helical, lipoyl, barrel, and membrane- proximal domains (MPD). It links inner membrane transporters and OMFs to form
Fig. 3. Schematic representation of bacterial eux pumps. All bacterial eux pumps are located on the inner membrane. Gram- negative bacteria
have three components in their cell envelope, i.e. outer membrane, peptidoglycan layer, and inner membrane. Gram- positive bacteria have only two
components in their cell envelope, i.e. peptidoglycan layer and inner membrane. Representative structures from each family (superfamily) have
been presented here. Currently, six types of eux pump families have been identified in bacteria, i.e. ATP- binding cassette (ABC) superfamily, major
facilitator superfamily (MFS), small multidrug resistance (SMR) family, proteobacterial antimicrobial compound eux (PACE) family, multidrug and
toxin extrusion (MATE) family, and resistance- nodulation- cell division (RND) superfamily. The following PDB (Protein Data Bank) identifiers were used
for depicting 3D structures; 2HYD for the ABC superfamily (Gram- positive); 3VVN for the MATE family; 4ZOW for MFS; 5NIK for the ABC superfamily
(Gram- negative); 5V5S for RND superfamily; and 6WK9 for SMR family. Three dimensional structure of the PACE family (GeneBank ID: A1S_2063) was
predicted using the ColabFold algorithm [119]. Then 3D structures were rendered using BIOVIA Discovery Studio Visualizer (Dassault Systems, France).

5
Gaurav etal., Microbiology 2023;169:001333
continuous conduits across the membrane [36, 37]. One of the most common enteric pathogens that aects both humans
and animals is S. enterica serovar Typhimurium. Consumption of contaminated food and water is the main source of bacteria
reaching the intestinal epithelial and it causes gastrointestinal disease [38]. Previous reports on S. enterica have described
that there are four three dimensional (3D) sites present in the promoter region of PAP sequence which are involved in the
structural foundation for PAP- RND interaction. According to structural studies, these four 3D sites correspond to nine
distinct linear binding sequences known as binding boxes [37]. Additionally, the aforementioned study identied critical
conserved residues within the binding boxes responsible for RND eux pump functionality and further rened the exact
residues of box4 (T271 and F292), box1 (R59) and box9 (K366) critical for eux function [37, 39]. e mutation of the
identied conserved residues of each binding box completely destabilises an RND- based tripartite eux pump, the AcrABZ-
TolC assembly. ese newly identied residues can be employed as novel therapeutic targets for the development of eux
pump inhibitors to combat antimicrobial resistance [37, 39]. RND superfamilies are known for their role in virulence as well
as resistance. e capacity for bacterial adhesion and invasion of host cells is crucial for eective colonisation and infection
[40]. A recent study found a signicant reduction in adhesion and invasion in eux pump mutants of E. coli in comparison
to parental strains, which is incongruent with a previous report where lack of AcrB in S. enterica reduced adhesion and
invasion, supporting a relevant role of the RND eux pump in bacterial virulence [41, 42].
Major facilitator superfamily (MFS)
e MFS family is the largest known superfamily of secondary active transporters found ubiquitously in bacteria, archaea,
and eukaryotes. It includes members that are solute uniport (movement of solute independent of ions), solute/cation symport
(movement of solute and ion in the same direction), solute/cation antiport (movement of ion and solute in opposite directions)
and/or solute/solute antiport with inwardly and/or outwardly directed polarity [43, 44]. Both symporters and antiporters
use energy from the proton motive force (PMF) to move substrates across the membrane [45]. e majority of this family
functions as a monomeric unit and possesses 12 to 14 TMHs which are organized into two domains, each as a bundle of six
helices [46]. Surprisingly, it is still unclear how MFS transporters are allosterically regulated down to the molecular level, but
with improvements in single- particle cryo- EM, and other techniques such as uorescence resonance energy transfer (FRET),
nuclear magnetic resonance (NMR) and electron paramagnetic resonance (EPR) spectroscopy, this can expand structural
data and reveal novel, conceptual insights into the MFS transporters [47, 48]. MFS transporters play important roles in
host–pathogen communication, especially in adhesion, invasion, intracellular survival, and biolm formation. Inhibiting
the activity of MDR transporters is a promising strategy to combat drug resistance and reduce virulence of pathogens, e.g.
inactivation of Acinetobacter baumannii MFS eux pump, AbaF, reduces bacterial virulence in a Caenorhabditis elegans
model [49, 50]. N. gonorrhoeae is a strict human pathogen that causes gonorrhoea, a sexually transmitted disease. FarAB,
an MFS pump in N. gonorrhoeae, mediates resistance to long- chained fatty acids like oleic, linoleic, and palmitic acids [51].
L. monocytogenes is a foodborne pathogen that can break the intestinal barrier, and rapidly spread to liver and spleen. Two
MFS pumps, MdrM and MdrT, play important roles in pathophysiology of L. monocytogenes. Both MdrM and MdrT are
involved in bile resistance and modulate the cytosolic surveillance pathway of innate immunity, which promotes bacterial
spread and tissue invasion. MdrM actively controls cytosolic bacteria’s ability to induce IFN-β expression [52].
Multidrug and toxic compound extrusion (MATE) family
e MATE family was rst reported 24 years ago and was believed to be closely related to the MFS family, although compu-
tational studies and structural characterization show that both the MATE and MFS families dier in their sequence and
topology [53–55]. MATE transporters are classied into the multidrug/oligosaccharide- lipid/polysaccharide (MOP) ippase
superfamily and further segregated into NorM, DNA damage- inducible protein F (DinF) and eukaryotic superfamilies based
on their amino acid sequence similarity [16, 56]. Several crystal structures of the MATE superfamily transporters have shown
that they consist of 12 transmembrane helices with an N- lobe and a C- lobe belonging to intramolecular pseudo twofold
symmetry, with an axis perpendicular to the membrane plane [55]. Despite structural similarities, the transport mechanisms
of a few MATE superfamily members have been observed to dier, for example, NorM- Vc from Vibrio cholerae and NorM- PS
from Pseudomonas stutzeri. NorM- VC and NorM- PS have comparable structural properties, however NorM- PS is driven
by H+ electrochemical gradients, while NorM- VC has been found to be coupled to both Na+ and H+ [57]. To translocate
cationic substrates across the membrane, the MATE superfamily functions as a secondary antiporter (inux of H+ or Na+)
and uses a rocker- switch alternating access mechanism, switching between substrate bound outward facing or ion- bound
inward- facing conformations [53]. Ethidium bromide, berberine, acriavine, noroxacin and tetraphenylphosphonium are
some of the cationic substrates that MATE transporters extrude and reduce susceptibility to these drugs in bacteria [58].
Bacteria are also prone to developing resistance following changes in the expression level of the MATE superfamily, for
example overexpression of MepA in Staphylococcus aureus can lead to resistance against tigecycline, which is used to treat
methicillin and vancomycin- resistant S. aureus infections [58]. e MATE superfamily eux (A1S_3371) pump has been
reported to contribute to A. baumannii ATCC 17978 virulence [59].

6
Gaurav etal., Microbiology 2023;169:001333
Small multidrug resistance (SMR) family
e SMR family is composed of small (12 kDa) integral inner membrane proteins containing only four transmembrane
α-helices, which confer resistance to a variety of quaternary ammonium compounds and other lipophilic cations in archaea
and bacteria [60]. Despite being small in size, the SMR family functions as homodimers or heterodimers. e overall
mechanism of SMR family for transport is an exchange between the substrate and a proton (antiport) [61]. A wide range of
SMR family studies have shown that multidrug resistance is driven by the proton motive force and that the conserved amino
acid glutamic acid is an important residue that contributes in expelling cationic drugs [62, 63]. Several SMR proteins have
been identied in bacterial pathogens and resistance has been found against clinically used antibiotics such as β-lactams,
aminoglycosides, inhibitors of dihydrofolate, and various antiseptics [64]. M. tuberculosis is a pathogenic bacterium and
the causative agent of tuberculosis that contains the mmr gene which encodes the Mmr protein. Overexpression of Mmr
protein reduces the bacterial susceptibility to ethidium bromide, quaternary ammonium compounds, and a few antibiotics
such as kanamycin and amikacin [65, 66]. As far as we know, little evidence has been shown for the contribution of MATE
transporters to bacterial virulence in human pathogens.
Proteobacterial antimicrobial compound eux (PACE) family
e PACE family proteins, with AceI from A. baumannii as the prototype, are a recently discovered family of bacterial drug
eux transport proteins that are encoded by genes in the bacterial core genome rather than by mobile genetic elements, which
suggests that they provide some important function [67]. In addition, AceI exhibits a wide range of resistance against struc-
turally diverse antimicrobial compounds and biosynthetic biocides (e.g. benzalkonium, diqualinium, acriavin, proavin,
and chlorhexidine) [68]. An alignment of 47 dierent PACE family proteins from various bacterial species revealed that
four amino acid residues, glutamic acid, asparagine, alanine, and aspartic acid, appeared to be conserved across the family.
Glutamic acid is found in transmembrane helix one, asparagine is found in transmembrane helix two, alanine is located at
the boundary of the periplasmic membrane of transmembrane helix four, and aspartic acid is located at the boundary of
the cytoplasmic membrane of transmembrane helix four [68]. AceI of A. baumannii is the prototype PACE family member
involved in the transport of widely used antiseptic – chlorhexidine. AceI’s glutamic acid (E15) was discovered to be fully
conserved and responsible for proton binding. In a recent study, E15 was found to play an important role in the dimerization
of AceI proteins in solution, where monomeric and dimeric forms of AceI proteins exist in a dynamic equilibrium and the
equilibrium state is modulated by pH, cardiolipin and chlorhexidine binding. Further, mutation of this glutamic acid into
glutamine (E15Q) results in a signicantly dierent pH response than the wild- type AceI protein [69]. e C- terminus of the
AceI protein was found to be highly conserved as compared to the N- terminus, which suggests that it plays an important role
in core function and, on the other hand, the N- terminus plays a major role in substrate recognition. Short- chain diamines
(such as cadaverine and putrescine) that are known for their important roles in metabolism, transcription regulation, and
protein expression were found to be physiological substrates of PACE family transporters [70]. PACE transporters translocate
their substrates using energy generated through the electrochemical proton gradient across the membrane, for instance the
PACE transporter PA2880 from P. aeruginosa mediates chlorhexidine eux by employing the proton motive force [71]. e
lack of genes encoding proteins of the PACE family in the E. coli chromosome also suggests that these genes were lost early
in the divergence from the Gammaproteobacteria [16].
STRATEGIES TO DISCOVER NOVEL EFFLUX PUMP INHIBITORS THAT CAN REVIVE THE ACTIVITY
OF INEFFECTIVE ANTIBIOTICS
Antibiotic resistance has grown parallelly whenever new antibiotics have been introduced to the market [6, 7]. Although
some antibiotic resistance determinants were present before the introduction of a particular antibiotic in the market; high
usage of antibiotics throughout the world exerts a selection pressure on bacteria and that can contribute to rise in antibiotic
resistance [4, 5, 72–74]. Eux pumps are major determinants of antibiotic resistance and hence identication of chemical
moieties capable of reversing the eect of eux pumps is needed. Several strategies have been employed to discover eux
pump inhibitors (EPIs) in the past. is section will focus on previously used strategies as well as the scope of introducing
emerging approaches like machine learning in discovery of EPIs.
Plant-derived secondary metabolites – a good starting point for EPIs
Combinatorial chemistry has changed how we think of developing a new chemical entity (NCE) however, plant- based natural
products have always been a source of human medicines. Natural products have been the single most productive source
for lead identication and further development of drugs. Plant metabolites have been actively used in drug development
for many therapeutic areas like cancer, neurological, cardiovascular, skin, gastrointestinal, inammation, and metabolic
disorders [75]. Additionally, many plant metabolites like terpenoids, phenolic compounds, and alkaloids show moderate
antibacterial activity [76]. e direct antibacterial eect of these secondary metabolites is due to their ability to hamper
bacterial protein synthesis, DNA synthesis, and RNA synthesis; however, their eect on the cell envelope is quite prominent

7
Gaurav etal., Microbiology 2023;169:001333
[77]. Plant secondary metabolites can cause severe bacterial cell envelope stress and that triggers a cascade of events such as
cell wall damage and the leakage of cytoplasmic constituents, metabolites and ions [76]. Although many intensive screen-
ings have been done to discover potent plant secondary metabolites as antibacterial agents, however, there is a scarcity of
literature describing the screening of plant secondary metabolites as eux pump inhibitors [78, 79]. Targeted screening of
plant secondary metabolites as EPIs is an important and currently overlooked strategy. Out of approximately 250 000 higher
plants worldwide, only about 14–28 % have been investigated for a medical purpose [79]. us, this inadequately explored
wealth of medicinal plants could be our ‘goldmine’ for the discovery of novel and eective eux pump inhibitors. In the
quest of nding novel antimicrobial entities during an antimicrobial screening of medicinal plants, we might have missed
many potential and potent EPIs because priority is usually given to plant metabolites and their derivatives showing evident
and signicant growth inhibition of test pathogens [80]. Here, we propose a new perspective that could help identify novel
EPIs from fractions of plant metabolites. An ideal EPI should not possess inherent antibacterial activity, as any chemical
entity exerting direct stress on bacterial physiology will suer the same consequence as of any antibiotics, i.e. the rapid
development of resistance and we have faced such outcomes with many antibiotics in the past [74].
e rise in the incidence of MDR phenotypes among pathogenic bacteria has consequently forced us to use combinations
of antibiotics. e depth of this crisis level can be understood by the fact that more than half (55%) of the combination
drugs approved by the Food and Drug Administration (FDA) are used to treat infections [81]. However, currently, there are
no approved EPIs for clinical usage. ere might be several factors for such a void; the rst is the diversity of eux pumps
available to any pathogen. Both Gram- positive and Gram- negative pathogens have multiple functional eux pumps. Addi-
tionally, broad substrate specicity is a hallmark of almost all eux pumps; it becomes even more complicated to develop
EPIs. Surprisingly, plants have developed multiple arsenals to tackle this problem. Most of the plant secondary metabolites
do not have inherent antibacterial activity except a few with moderate antibacterial activity. Indeed, many of them act as
antiherbivoral and other molecules required for interspecies competition [80, 82].
Interestingly, plants have developed many eux pump inhibitors that act synergistically with other secondary metabolites
like alkaloids and thus making a non- antibacterial secondary metabolite or moderately active secondary metabolite into a
potent antibacterial [80]. For example, 5′-methoxyhydnocarpin- D (5′-MHC- D) is a avonolignan produced by species like
Berberis fremontii, Berberis repens, and Berberis aquifolia – berberine- producing native American plants. 5′-MHC- D has no
antibacterial activity on its own. However, it acts as a very potent eux pump inhibitor making berberine as eective as or
even better than clinically used antibiotics against multiple drug- resistant S. aureus [83–85]. e overall idea is that the plants
deploy a range of phytochemicals to potentiate a range of weak antimicrobials or narrow- spectrum antimicrobials to ght
against all sets of invading bacterial pathogens. Hence, in order to harness the power of plant natural product diversity, we
need to change our screening strategy. Since most of the plant diversity is unexplored for screening eux pump inhibitors,
there is an urgent need to look for new screening strategies that would yield novel eux pump inhibitors. We need to collect
plants from every corner of the world and a parallel bioactive compound extraction method could be followed. Previously
reported weak and narrow- spectrum antibacterial phytochemicals as well as currently used antibiotics could be retested
along with newly extracted phytochemicals. e major issue with the current eux pump inhibitor screening is the lack of
diversity; most of the screenings reported previously have been performed using recombinant strains expressing a single
eux pump like NorA, AcrAB‐TolC, MexAB- OprM or AbeM [84, 86–88]. e problem with the leads obtained from current
screenings is that they inhibit only specic eux pumps, however in clinical settings pathogens are equipped with eux
pumps from multiple classes [89–92]. ese scenarios make discovering a clinically feasible eux pump inhibitor an arduous
task. In order to broaden the therapeutic range of an eux pump inhibitor, screening against recombinant strains expressing
multiple eux pumps prevalent among clinical strains is essential (Fig.4). Although the percentage of positive hits in such
screening will reduce signicantly due to stringent criteria, this will yield an enhanced condence score. Structure- activity
relationship studies of previously identied eux pump inhibitors from dierent screening programmes may provide valuable
information in this avenue [87, 88].
Conjugation of an eux pump inhibitor to a weak antibacterial molecule is also a good strategy to start with [93, 94]. Berberine
is a weak cationic antibacterial molecule that is prone to euxed out by multiple eux pumps in dierent pathogens. Surpris-
ingly, conjugation of INF55 (an eux pump inhibitor having no inherent antibacterial activity) with berberine yielded a potent
antibacterial molecule with good in vivo ecacy against deadly enterococcal infection [94].
Eux pump inhibitors from the past
To date, several potent eux pump inhibitors have been discovered. ese eux pump inhibitors belong to structurally
diverse chemical classes like peptidomimetics [95], piperazines [96], pyridopyrimidines [97], and pyranopyrimidines [98].
MC- 207,110 (Phe- Arg-β-naphthylamide or PAβN) is a broad- spectrum eux pump inhibitor from peptidomimetics class
that was identied using large scale screening of a chemical library [95]. Next, 1- (1- naphthylmethyl)- piperazine (NMP)
is a potent eux pump inhibitor from piperazines class specically active against RND class eux pumps [96]. D13- 9001
is a potent and safe eux pump inhibitor from pyridopyrimidines class which shows exceptional in vivo activity against

8
Gaurav etal., Microbiology 2023;169:001333
P. aeruginosa infection [97]. MBX2319 is a recently discovered eux pump inhibitor of AcrAB- TolC that can potentiate
multiple antibiotics such as levooxacin, piperacillin, and cefotaxime. Recently, two new eux pump inhibitors (chlor-
promazine and amitriptyline) have been identied using drug repurposing platforms. Both of these drugs have antipsychotic
properties [99]. However, none of the above- mentioned molecules have been approved as eux pump inhibitors for clinical
usage basically due to in vivo cytotoxicity and other adverse eects [100].
Machine learning-based discovery of novel eux pump inhibitors
e domain of chemoinformatics has changed over the past few years. ere are several models for molecular property
prediction of any chemical entity. Molecular ngerprinting is one of them and it is an essential cheminformatics tool for
virtual screening and mapping chemical space. For small molecules, substructure ngerprints are the preferred technique,
while for large molecules (e.g. antimicrobial peptides), atom- pair ngerprints are preferable. Still, there is no common
method that achieves good performance on both classes of molecules [101]. Traditionally, molecules were represented
by their molecular ngerprint vectors based on the presence or absence of functional groups in the molecule (Fig.5a)
[102]. e second strategy is to use molecular descriptors that are based on molecular properties and needs supervision by
domain experts (Fig.5b) [103]. ese drawbacks and the limited accuracy of these models restrict their usage to small sets
of molecules (a few thousand) [104]. However, recent innovations in neural network algorithms provide an opportunity
to inuence the paradigm of antibacterial drug discovery [104, 105]. Now any chemical entity (for example eux pump
inhibitor) can be denoted by a hybrid representation that contains both convolutions and molecular descriptors. is new
algorithm outperforms any previous one by providing exibility in learning a task around a xed molecular descriptor (like
a pharmacophore), simultaneously making convolutions around bonds rather than atoms. is last step also helps to reduce
the total running time by averting unnecessary loops during the message- passing phase in the algorithm [105].
Advancements in high- performance computing and parallel computing along with the availability of better computer hard-
ware makes it a perfect time to implement machine learning for nding new antibacterial entities. Machine learning is a
powerful tool that can be used to generate predictive as well as generative models which can help us to ght against bacterial
infections. In recent times, several research groups have deployed machine learning approach to explore new antibacterials
[104, 106–108]. Machine learning and computational approaches have helped to design novel antimicrobial peptides (AMPs)
with enhanced ecacy and reduced toxicity [107]. Furthermore, the search for novel and cryptic antimicrobial peptides
in the human genome and the human microbiome has yielded some interesting leads with excellent antibacterial activity
[106, 108]. Recently, a deep learning approach has helped to discover an unusual class of antibiotic - Halicin, which has a
broad- spectrum antibacterial activity [104]. One of the most important aspects of machine learning is the avoidance of the
dereplication problem, wherein the same molecules are repeatedly discovered. e machine learning approach could lter out
already reported antibacterial molecules. Recent advancements in the eld of machine learning algorithms could denitely
help us predict molecular properties of novel eux pump inhibitors (Fig.6). We can also capture vast chemical spaces in
Fig. 4. Schematic representation describing an approach to discover new eux pump inhibitors (EPIs) from the natural products library. Screening
of the natural products library could help identify novel EPIs or a group of related chemical moieties that act synergistically as a potent antibacterial
agent. Screening for EPIs could be performed against recombinant strains expressing multiple eux pumps to select a broad- spectrum EPI or
collection of EPIs that can target dierent eux pumps.

9
Gaurav etal., Microbiology 2023;169:001333
silico that are beyond the reach of the current experimental approach with far less associated running cost. Until now, search
for novel eux pump inhibitors using a machine learning approach has not been initiated (zero results on PubMed, searched
on 24 November 2022). It is now more important than ever to screen for eux pump inhibitors capable of restoring the
eectiveness of ‘magic bullets’ using a deep learning approach. is can be achieved in three stages; rst, the collection of
molecular descriptors of positive hits from previously screened chemical libraries can be taken up [88, 95, 98, 109–112].
Second, training of a deep neural network(s) model to predict growth inhibition of test pathogen(s) using a combination
of potential eux pump inhibitor (positive hits from the previously screened chemical library) and test antibiotic can be
Fig. 5. (a)Three categories of molecular descriptors; 1D (one- dimensional), 2D, and 3D descriptors. 1D descriptors depend on molecular formula; 2D
descriptors contain 2D molecular fingerprints; 3D descriptors provide information about 3D geometric information of any molecule. (b)Schematic
representation describing the process of making a graph convolution model of a molecule. G represents graph descriptors, A represents a set of atoms,
B represents a set of bonds, and X and Y represent the atom content matrix.
Fig. 6. Schematic representation describing an approach to discover new eux pump inhibitors (EPIs) using machine learning. The machine learning
approach could help identify novel and robust EPIs using information already available for existing EPIs. Machine learning algorithms extract common
feature among training data sets and implement them to find a lead among testing datasets.

10
Gaurav etal., Microbiology 2023;169:001333
taken up. ird, applying the best model to several discrete chemical libraries (with more than 100 million compounds) like
ZINC15 and Maybridge to identify potential lead eux pump inhibitors [113] (Fig.6).
Challenges and perspectives of bacterial eux pump inhibitors
There are a few important concerns that need to be discussed before developing broad- spectrum efflux pump inhibitors.
First, efflux pumps provide protection only in actively growing bacterial cells [114]. This could limit their potential usage
against slow- growing or nongrowing pathogens where reduced membrane permeability plays a major role in maintaining
low antibiotic concentration. Second, efflux pumps are not the one and only mechanism behind antibiotic resistance in
bacteria. Third, all efflux pumps are not exclusive to the bacterial kingdom. Broad- spectrum ATP- dependent efflux pumps
are also present in humans. MDR cancer cells often overexpress ATP- dependent efflux pumps to avoid toxicity caused by
anti- cancer agents. A broad- spectrum efflux pump inhibitor may also target efflux pumps present on human cells and thus
may show side- effects. Fourth, by virtue of their nature, efflux pump inhibitors will be used in combination with partner
antibiotics. This poses an additional challenge to appropriately tailor the pharmacokinetic properties of both components
of the combinations.
us many challenges are encountered on the path to conversion of a drug lead to a clinically valuable therapeutic agent.
In this regard, an eux pump inhibitor is no dierent from any other new chemical entity. However, what if we can hunt
eux pump inhibitors among existing drugs already used clinically for a dierent indication. is could signicantly reduce
time and resources for regulatory approval [115]. Similarly, many transporters are found in representatives of all kingdoms.
However, homology between bacterial and human proteins is negligibly low and even if there are sequence similarities,
conserved regions might be located in an integral inner membrane protein that does not participate in substrate specicity
[116–118].
CONCLUSION
Bacterial eux systems play a crucial role in antibiotic resistance. ey contribute to intrinsic as well as acquired antibiotic
resistance in bacteria. e bacterial eux system aects virtually all classes of antibiotics. Eux pump inhibitors capable of
restoring the eectiveness of available antibiotics are urgently needed. Untapped natural products can be a great resource
for potential eux pump inhibitors. Additionally, the machine learning approach can denitely help us to screen for new
eux pump inhibitors.
Funding information
This work was supported by a grant received from DBT- Wellcome Trust India Alliance (Grant number: TIA- 1815- BIO) to R.P. A.G. was supported by a
fellowship from the Department of Biotechnology, Government of India.
Acknowledgements
The authors would like to thank Harpreet Singh for insightful comments and critically reading the manuscript.
Conflicts of interest
The authors declare no competing interests.
References
1. Laxminarayan R, Duse A, Wattal C, Zaidi AKM, Wertheim HFL,
etal. Antibiotic resistance- the need for global solutions. Lancet
Infect Dis 2013;13:1057–1098.
2. Lewis K. The science of antibiotic discovery. Cell 2020;181:29–45.
3. CDC. Antibiotic Resistance Threats in the United States; 2019
4. D’Costa VM, King CE, Kalan L, Morar M, Sung WWL, et al. Antibi-
otic resistance is ancient. Nature 2011;477:457–461.
5. Larsen J, Raisen CL, Ba X, Sadgrove NJ, Padilla- González GF,
et al. Emergence of methicillin resistance predates the clinical
use of antibiotics. Nature 2022;602:135–141.
6. WHO. Antimicrobial Resistance; 2021. https://www.who.int/
news-room/fact-sheets/detail/antibiotic-resistance [accessed
15 March 2023].
7. Van Boeckel TP, Brower C, Gilbert M, Grenfell BT, Levin SA, etal.
Global trends in antimicrobial use in food animals. Proc Natl Acad
Sci 2015;112:5649–5654.
8. WHO. Global priority list of antibiotic- resistant bacteria to guide
research, discovery, and development of new antibiotics. pdf;
2017
9. Piddock LJV. Multidrug- resistance eux pumps - not just for
resistance. Nat Rev Microbiol 2006;4:629–636.
10. Webber MA, Piddock LJV. The importance of eux pumps
in bacterial antibiotic resistance. J Antimicrob Chemother
2003;51:9–11.
11. Poole K. Eux- mediated resistance to fluoroquinolones in gram-
positive bacteria and the mycobacteria. Antimicrob Agents Chem-
other 2000;44:2595–2599.
12. Piddock LJV. Clinically relevant chromosomally encoded multi-
drug resistance eux pumps in bacteria. Clin Microbiol Rev
2006;19:382–402.
13. Bazzi W, Abou Fayad AG, Nasser A, Haraoui L- P, Dewachi O,
et al. Heavy metal toxicity in armed conflicts potentiates AMR
in A. baumannii by selecting for antibiotic and heavy metal
co- resistance mechanisms. Front Microbiol 2020;11:68.
14. Gaurav A, Sharma A, Pathania R. Eux Mediated Co- resistance. In:
Mandal SM, Paul D (editors). Bacterial Adaptation to Co- Resistance.
Singapore: Springer Singapore; 2019. pp. 161-180.
15. Nies DH. Eux- mediated heavy metal resistance in prokaryotes.
FEMS Microbiol Rev 2003;27:313–339.

11
Gaurav etal., Microbiology 2023;169:001333
16. Du D, Wang- Kan X, Neuberger A, van Veen HW, Pos KM, et al.
Multidrug eux pumps: structure, function and regulation. Nat
Rev Microbiol 2018;16:523–539.
17. Sharma A, Gupta VK, Pathania R. Eux pump inhibitors for
bacterial pathogens: From bench to bedside. Indian J Med Res
2019;149:129.
18. Lubelski J, Konings WN, Driessen AJM. Distribution and physi-
ology of ABC- type transporters contributing to multidrug resist-
ance in bacteria. Microbiol Mol Biol Rev 2007;71:463–476.
19. Dean M, Moitra K, Allikmets R. The human ATP- binding cassette
(ABC) transporter superfamily. Hum Mutat 2022;43:1162–1182.
20. Rees DC, Johnson E, Lewinson O. ABC transporters: the power to
change. Nat Rev Mol Cell Biol 2009;10:218–227.
21. Thomas C, Tampé R. Structural and mechanistic principles of
ABC transporters. Annu Rev Biochem 2020;89:605–636.
22. Tarling EJ, Vallim TQ de A, Edwards PA. Role of ABC transporters
in lipid transport and human disease. Trends in Endocrinol Metab
2013;24:342–350.
23. Javed W, Vallet S, Clement M- P, Le Roy A, Moulin M, etal. Struc-
tural insights into the catalytic cycle of a bacterial multidrug ABC
eux pump. J Mol Biol 2022;434:167541.
24. Akhtar AA, Turner DP. The role of bacterial ATP- binding cassette
(ABC) transporters in pathogenesis and virulence: therapeutic
and vaccine potential. Microb Pathog 2022;171:105734.
25. Bogomolnaya LM, Andrews KD, Talamantes M, Maple
A, Ragoza Y, et al. The ABC- type eux pump MacAB protects
Salmonella enterica serovar typhimurium from oxidative stress.
mBio 2013;4:e00630–13.
26. Palmer LD, Skaar EP. Transition metals and virulence in bacteria.
Annu Rev Genet 2016;50:67–91.
27. Schwan WR, Warrener P, Keunz E, Stover CK, Folger KR. Muta-
tions in the cueA gene encoding a copper homeostasis P- type
ATPase reduce the pathogenicity of Pseudomonas aeruginosa in
mice. Int J Med Microbiol 2005;295:237–242.
28. Francis MS, Thomas CJ. Mutants in the CtpA copper transporting
P- type ATPase reduce virulence of Listeria monocytogenes.
Microb Pathog 1997;22:67–78.
29. Ward SK, Abomoelak B, Hoye EA, Steinberg H, Talaat AM. CtpV: a
putative copper exporter required for full virulence of Mycobacte-
rium tuberculosis. Mol Microbiol 2010;77:1096–1110.
30. Nikaido H. RND transporters in the living world. Res Microbiol
2018;169:363–371.
31. Breyton C, Haase W, Rapoport TA, Kühlbrandt W, Collinson I.
Three- dimensional structure of the bacterial protein-
translocation complex SecYEG. Nature 2002;418:662–665.
32. Klenotic PA, Moseng MA, Morgan CE, Yu EW. Structural and
functional diversity of resistance- nodulation- cell division trans-
porters. Chem Rev 2021;121:5378–5416.
33. Tseng TT, Gratwick KS, Kollman J, Park D, Nies DH, etal. The RND
permease superfamily: an ancient, ubiquitous and diverse family
that includes human disease and development proteins. J Mol
Microbiol Biotechnol 1999;1:107–125.
34. Neuberger A, Du D, Luisi BF. Structure and mechanism of bacte-
rial tripartite eux pumps. Res Microbiol 2018;169:401–413.
35. Alav I, Kobylka J, Kuth MS, Pos KM, Picard M, et al. Structure,
assembly, and function of tripartite eux and type 1 secretion
systems in gram- negative bacteria. Chem Rev 2021;121:5479–5596.
36. Symmons MF, Marshall RL, Bavro VN. Architecture and roles
of periplasmic adaptor proteins in tripartite eux assemblies.
Front Microbiol 2015;6:513.
37. McNeil HE, Alav I, Torres RC, Rossiter AE, Laycock E, etal. Iden-
tification of binding residues between Periplasmic Adapter
Protein (PAP) and RND eux pumps explains PAP- pump
promiscuity and roles in antimicrobial resistance. PLoS Pathog
2019;15:e1008101.
38. Fàbrega A, Vila J. Salmonella enterica serovar Typhimurium skills
to succeed in the host: virulence and regulation. Clin Microbiol
Rev 2013;26:308–341.
39. Alav I, Bavro VN, Blair JMA. A role for the periplasmic adaptor
protein AcrA in vetting substrate access to the RND eux trans-
porter AcrB. Sci Rep 2022;12:4752.
40. Spaniol V, Bernhard S, Aebi C. Moraxella catarrhalis AcrAB-
OprM eux pump contributes to antimicrobial resistance and is
enhanced during cold shock response. Antimicrob Agents Chem-
other 2015;59:1886–1894.
41. Wang- Kan X, Blair JMA, Chirullo B, Betts J, La Ragione RM, etal.
Lack of AcrB eux function confers loss of virulence on Salmo-
nella enterica serovar Typhimurium . mBio 2017;8:e00968–00917.
42. Fanelli G, Pasqua M, Prosseda G, Grossi M, Colonna B. AcrAB
eux pump impacts on the survival of adherent- invasive Escher-
ichia coli strain LF82 inside macrophages. Sci Rep 2023;13.
43. Yan N. Structural advances for the major facilitator superfamily
(MFS) transporters. Trends Biochem Sci 2013;38:151–159.
44. Pao SS, Paulsen IT, Saier MH. Major facilitator superfamily.
Microbiol Mol Biol Rev 1998;62:1–34.
45. Yan N. Structural biology of the major facilitator superfamily
transporters. Annu Rev Biophys 2015;44:257–283.
46. Radestock S, Forrest LR. The alternating- access mechanism of
MFS transporters arises from inverted- topology repeats. J Mol
Biol 2011;407:698–715.
47. Drew D, North RA, Nagarathinam K, Tanabe M. Structures and
general transport mechanisms by the Major Facilitator Super-
family (MFS). Chem Rev 2021;121:5289–5335.
48. Zhai G, Zhang Z, Dong C. Mutagenesis and functional analysis of
SotB: a multidrug transporter of the major facilitator superfamily
from Escherichia coli. Front Microbiol 2022;13:1024639.
49. Pasqua M, Bonaccorsi di Patti MC, Fanelli G, Utsumi R, Eguchi Y,
etal. Host - bacterial pathogen communication: the wily role of
the multidrug eux pumps of the MFS family. Front Mol Biosci
2021;8:723274.
50. Sharma A, Sharma R, Bhattacharyya T, Bhando T, Pathania R.
Fosfomycin resistance in Acinetobacter baumannii is mediated by
eux through a major facilitator superfamily (MFS) transporter-
AbaF. J Antimicrob Chemother 2017;72:68–74.
51. Lee EH, Shafer WM. The farAB- encoded eux pump mediates
resistance of gonococci to long- chained antibacterial fatty acids.
Mol Microbiol 1999;33:839–845.
52. Crimmins GT, Herskovits AA, Rehder K, Sivick KE, Lauer P, etal.
Listeria monocytogenes multidrug resistance transporters acti-
vate a cytosolic surveillance pathway of innate immunity. Proc
Natl Acad Sci 2008;105:10191–10196.
53. Morita Y, Kodama K, Shiota S, Mine T, Kataoka A, etal. NorM,
a putative multidrug eux protein, of Vibrio parahaemolyticus
and its homolog in Escherichia coli. Antimicrob Agents Chemother
1998;42:1778–1782.
54. Brown MH, Paulsen IT, Skurray RA. The multidrug eux protein
NorM is a prototype of a new family of transporters. Mol Microbiol
1999;31:394–395.
55. Miyauchi H, Moriyama S, Kusakizako T, Kumazaki K, Nakane T,
etal. Structural basis for xenobiotic extrusion by eukaryotic MATE
transporter. Nat Commun 2017;8:1633.
56. Hvorup RN, Winnen B, Chang AB, Jiang Y, Zhou X- F, et al. The
multidrug/oligosaccharidyl- lipid/polysaccharide (MOP) exporter
superfamily. Eur J Biochem 2003;270:799–813.
57. Stephen J, Lekshmi M, Ammini P, Kumar SH, Varela MF.
Membrane eux pumps of pathogenic Vibrio species: role in anti-
microbial resistance and virulence. Microorganisms 2022;10:382.
58. Kim J, Cater RJ, Choy BC, Mancia F. Structural insights into
transporter- mediated drug resistance in infectious diseases. J
Mol Biol 2021;433:167005.
59. Pérez- Varela M, Corral J, Aranda J, Barbé J. Roles of eux
pumps from dierent superfamilies in the surface- associated

12
Gaurav etal., Microbiology 2023;169:001333
motility and virulence of Acinetobacter baumannii ATCC 17978.
Antimicrob Agents Chemother 2019;63:e02190–02118.
60. Paulsen IT, Skurray RA, Tam R, Saier MH Jr, Turner RJ, et al. The
SMR family: a novel family of multidrug eux proteins involved
with the eux of lipophilic drugs. Mol Microbiol 1996;19:1167–1175.
61. Van Bambeke F, Balzi E, Tulkens PM. Antibiotic eux pumps.
Biochem Pharmacol 2000;60:457–470.
62. Gottschalk KE, Soskine M, Schuldiner S, Kessler H. A structural
model of EmrE, A multi- drug transporter from Escherichia coli.
Biophys J 2004;86:3335–3348.
63. Chung YJ, Saier MH. Overexpression of the Escherichia coli sugE
gene confers resistance to a narrow range of quaternary ammo-
nium compounds. J Bacteriol 2002;184:2543–2545.
64. Bay DC, Rommens KL, Turner RJ. Small multidrug resistance
proteins: a multidrug transporter family that continues to grow.
Biochim Biophys Acta 2008;1778:1814–1838.
65. Rodrigues L, Villellas C, Bailo R, Viveiros M, Aínsa JA. Role of the
Mmr eux pump in drug resistance in Mycobacterium tubercu-
losis. Antimicrob Agents Chemother 2013;57:751–757.
66. Balganesh M, Dinesh N, Sharma S, Kuruppath S, Nair AV, etal.
Eux pumps of Mycobacterium tuberculosis play a significant
role in antituberculosis activity of potential drug candidates. Anti-
microb Agents Chemother 2012;56:2643–2651.
67. Hassan KA, Liu Q, Henderson PJF, Paulsen IT. Homologs of the
Acinetobacter baumannii AceI transporter represent a new family
of bacterial multidrug eux systems. mBio 2015;6:e01982- 14.
68. Hassan KA, Liu Q, Elbourne LDH, Ahmad I, Sharples D, et al.
Pacing across the membrane: the novel PACE family of eux
pumps is widespread in Gram- negative pathogens. Res Microbiol
2018;169:450–454.
69. Bolla JR, Howes AC, Fiorentino F, Robinson CV. Assembly and
regulation of the chlorhexidine- specific eux pump AceI. Proc
Natl Acad Sci USA 2020;117:17011–17018.
70. Hassan KA, Naidu V, Edgerton JR, Mettrick KA, Liu Q, etal. Short-
chain diamines are the physiological substrates of PACE family
eux pumps. Proc Natl Acad Sci 2019;116:18015–18020.
71. Zhao J, Hellwig N, Djahanschiri B, Khera R, Morgner N, et al.
Assembly and functional role of PACE transporter PA2880 from
Pseudomonas aeruginosa. Microbiol Spectr 2022;10:e0145321.
72. Gould IM. Coping with antibiotic resistance: the impending crisis.
Int J Antimicrob Agents 2010;36 Suppl 3:S1–2.
73. Brandt C, Makarewicz O, Fischer T, Stein C, Pfeifer Y, et al. The
bigger picture: the history of antibiotics and antimicrobial resist-
ance displayed by scientometric data. Int J Antimicrob Agents
2014;44:424–430.
74. Davies J, Davies D. Origins and evolution of antibiotic resistance.
Microbiol Mol Biol Rev 2010;74:417–433.
75. Harvey AL. Natural products in drug discovery. Drug Discov Today
2008;13:894–901.
76. Burt S. Essential oils: their antibacterial properties and potential
applications in foods--a review. Int J Food Microbiol 2004;94:223–253.
77. Dhyani R, Srivastava SK, Shankar K, Ghosh T, Beniwal A,
etal. A chemical genetic approach using genetically encoded
reporters to detect and assess the toxicity of plant secondary
metabolites against bacterial pathogens. J Hazard Mater
2021;418:126399.
78. Li J- H, Vederas JC. Drug discovery and natural products: end of
an era or an endless frontier? Science 2009;325:161–165.
79. Seukep AJ, Kuete V, Nahar L, Sarker SD, Guo M. Plant-
derived secondary metabolites as the main source of eux
pump inhibitors and methods for identification. J Pharm Anal
2020;10:277–290.
80. Lewis K. In search of natural substrates and inhibitors of MDR
pumps. J Mol Microbiol Biotechnol 2001;3:247–254.
81. Sun X, Vilar S, Tatonetti NP. High- throughput methods for combi-
natorial drug discovery. Sci Transl Med 2013;5:205.
82. Mithöfer A, Boland W. Plant defense against herbivores: chem-
ical aspects. Annu Rev Plant Biol 2012;63:431–450.
83. Stermitz FR, Lorenz P, Tawara JN, Zenewicz LA, Lewis K. Synergy
in a medicinal plant: antimicrobial action of berberine potenti-
ated by 5’-methoxyhydnocarpin, a multidrug pump inhibitor. Proc
Natl Acad Sci 2000;97:1433–1437.
84. Stermitz FR, Tawara- Matsuda J, Lorenz P, Mueller P,
Zenewicz L, et al. 5’-Methoxyhydnocarpin- D and pheophorbide
A: Berberis species components that potentiate berberine
growth inhibition of resistant Staphylococcus aureus. J Nat Prod
2000;63:1146–1149.
85. Guz NR, Stermitz FR, Johnson JB, Beeson TD, Willen S, et al.
Flavonolignan and flavone inhibitors of a Staphylococcus aureus
multidrug resistance pump: structure- activity relationships. J
Med Chem 2001;44:261–268.
86. Hirakata Y, Kondo A, Hoshino K, Yano H, Arai K, etal. Eux pump
inhibitors reduce the invasiveness of Pseudomonas aeruginosa.
Int J Antimicrob Agents 2009;34:343–346.
87. Markham PN, Westhaus E, Klyachko K, Johnson ME,
Neyfakh AA. Multiple novel inhibitors of the NorA multidrug
transporter of Staphylococcus aureus. Antimicrob Agents Chem-
other 1999;43:2404–2408.
88. Bhattacharyya T, Sharma A, Akhter J, Pathania R. The
small molecule IITR08027 restores the antibacterial activity
of fluoroquinolones against multidrug- resistant Acineto-
bacter baumannii by eux inhibition. Int J Antimicrob Agents
2017;50:219–226.
89. Blair JMA, Webber MA, Baylay AJ, Ogbolu DO, Piddock LJV.
Molecular mechanisms of antibiotic resistance. Nat Rev Microbiol
2015;13:42–51.
90. Darby EM, Trampari E, Siasat P, Gaya MS, Alav I, etal. Molecular
mechanisms of antibiotic resistance revisited. Nat Rev Microbiol
2023;21:280–295.
91. Blair JMA, Richmond GE, Piddock LJV. Multidrug eux pumps
in Gram- negative bacteria and their role in antibiotic resistance.
Future Microbiol 2014;9:1165–1177.
92. Schindler BD, Kaatz GW. Multidrug eux pumps of Gram-
positive bacteria. Drug Resist Updat 2016;27:1–13.
93. Rineh A, Dolla NK, Ball AR, Magana M, Bremner JB, et al.
Attaching the NorA eux pump inhibitor INF55 to methylene blue
enhances antimicrobial photodynamic inactivation of methicillin-
resistant Staphylococcus aureus in vitro and in vivo. ACS Infect Dis
2017;3:756–766.
94. Ball AR, Casadei G, Samosorn S, Bremner JB, Ausubel FM,
etal. Conjugating berberine to a multidrug eux pump inhibitor
creates an eective antimicrobial. ACS Chem Biol 2006;1:594–600.
95. Lomovskaya O, Warren MS, Lee A, Galazzo J, Fronko R, et al.
Identification and characterization of inhibitors of multidrug
resistance eux pumps in Pseudomonas aeruginosa: novel
agents for combination therapy. Antimicrob Agents Chemother
2001;45:105–116.
96. Kern WV, Steinke P, Schumacher A, Schuster S, Baum H von,
et al. Eect of 1- (1- naphthylmethyl)- piperazine, a novel puta-
tive eux pump inhibitor, on antimicrobial drug susceptibility
in clinical isolates of Escherichia coli. J Antimicrob Chemother
2006;57:339–343.
97. Yoshida K, Nakayama K, Ohtsuka M, Kuru N, Yokomizo Y, et al.
MexAB- OprM specific eux pump inhibitors in Pseudomonas
aeruginosa. Part 7: highly soluble and in vivo active quaternary
ammonium analogue D13- 9001, a potential preclinical candi-
date. Bioorg Med Chem 2007;15:7087–7097.
98. Opperman TJ, Kwasny SM, Kim H- S, Nguyen ST, Houseweart C,
etal. Characterization of a novel pyranopyridine inhibitor of the
AcrAB eux pump of Escherichia coli. Antimicrob Agents Chem-
other 2014;58:722–733.
99. Grimsey EM, Fais C, Marshall RL, Ricci V, Ciusa ML, etal. Chlor-
promazine and amitriptyline are substrates and inhibitors of the
AcrB multidrug eux pump. mBio 2020;11:e00465- 20.

13
Gaurav etal., Microbiology 2023;169:001333
100. Li X- Z, Plésiat P, Nikaido H. The challenge of eux- mediated
antibiotic resistance in Gram- negative bacteria. Clin Microbiol Rev
2015;28:337–418.
101. Hou T, Wang J. Structure- ADME relationship: still a long way to
go? Expert Opin Drug Metab Toxicol 2008;4:759–770.
102. Rogers D, Hahn M. Extended- connectivity fingerprints. J Chem Inf
Model 2010;50:742–754.
103. Mauri A, Consonni V, Pavan M, Todeschini R. Dragon software:
an easy approach to molecular descriptor calculations. Match
2006;56:237–248.
104. Stokes JM, Yang K, Swanson K, Jin W, Cubillos- Ruiz A, etal. A deep
learning approach to antibiotic discovery. Cell 2020;180:688–702..
105. Yang K, Swanson K, Jin W, Coley C, Eiden P, et al. Analyzing
learned molecular representations for property prediction. J
Chem Inf Model 2019;59:3370–3388.
106. Torres MDT, Melo MCR, Flowers L, Crescenzi O, Notomista E,
etal. Author Correction: mining for encrypted peptide antibiotics
in the human proteome. Nat Biomed Eng 2022;6:1451.
107. Das P, Sercu T, Wadhawan K, Padhi I, Gehrmann S, et al. Accel-
erated antimicrobial discovery via deep generative models and
molecular dynamics simulations. Nat Biomed Eng 2021;5:613–623.
108. Porto WF, Irazazabal L, Alves ESF, Ribeiro SM, Matos CO, et al. In
silico optimization of a guava antimicrobial peptide enables combi-
natorial exploration for peptide design. Nat Commun 2018;9:1490.
109. Kumar A, Khan IA, Koul S, Koul JL, Taneja SC, et al. Novel
structural analogues of piperine as inhibitors of the NorA
eux pump of Staphylococcus aureus. J Antimicrob Chemother
2008;61:1270–1276.
110. Renau TE, Léger R, Flamme EM, Sangalang J, She MW, et al.
Inhibitors of eux pumps in Pseudomonas aeruginosa potentiate
the activity of the fluoroquinolone antibacterial levofloxacin. J
Med Chem 1999;42:4928–4931.
111. Bohnert JA, Kern WV. Selected arylpiperazines are capable of
reversing multidrug resistance in Escherichia coli overexpressing
RND eux pumps. Antimicrob Agents Chemother 2005;49:849–852.
112. Nakayama K, Ishida Y, Ohtsuka M, Kawato H, Yoshida K, et al.
MexAB- OprM- Specific eux pump inhibitors in Pseudomonas
aeruginosa. Part 1: Discovery and early strategies for lead opti-
mization. Bioorg Med Chem Lett 2003;13:4201–4204.
113. Sterling T, Irwin JJ. ZINC 15 – ligand discovery for everyone. J
Chem Inf Model 2015;55:2324–2337.
114. Whittle EE, McNeil HE, Trampari E, Webber M, Overton TW,
et al. Eux impacts intracellular accumulation only in actively
growing bacterial cells. mBio 2021;12.
115. Lomovskaya O, Bostian KA. Practical applications and feasi-
bility of eux pump inhibitors in the clinic—A vision for
applied use. Biochem Pharmacol 2006;71:910–918.
116. Tseng TT, Gratwick KS, Kollman J, Park D, Nies DH, et al.
The RND permease superfamily: an ancient, ubiquitous and
diverse family that includes human disease and development
proteins. J Mol Microbiol Biotechnol 1999;1:107–125.
117. Higgins CF. ABC transporters: from microorganisms to man.
Annu Rev Cell Biol 1992;8:67–113.
118. Dassa E, Hofnung M. Sequence of gene malG in E. coli K12:
homologies between integral membrane components from
binding protein- dependent transport systems. EMBO J
1985;4:2287–2293.
119. Mirdita M, Schütze K, Moriwaki Y, Heo L, Ovchinnikov S,
etal. ColabFold: making protein folding accessible to all. Nat
Methods 2022;19:679–682.
Five reasons to publish your next article with a Microbiology Society journal
1. When you submit to our journals, you are supporting Society activities for your community.
2. Experience a fair, transparent process and critical, constructive review.
3. If you are at a Publish and Read institution, you’ll enjoy the benefits of Open Access across
our journal portfolio.
4. Author feedback says our Editors are ‘thorough and fair’ and ‘patient and caring’.
5. Increase your reach and impact and share your research more widely.
Find out more and submit your article at microbiologyresearch.org.